22:59:12 #startmeeting Network team meeting, 23 January 2020 22:59:12 Meeting started Wed Jan 22 22:59:12 2020 UTC. The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot. 22:59:12 Useful Commands: #action #agreed #help #info #idea #link #topic. 22:59:16 hi! 22:59:17 hello network-team 22:59:34 who is here for our meeting today? :-) 22:59:43 our pad is at https://pad.riseup.net/p/tor-netteam-2020.1-keep 22:59:56 hi! 23:00:07 hello 23:00:22 catalyst, mikeperry, asn, dgoulet: are you joining us tonight? 23:00:25 o/ 23:00:26 and teor maybe 23:00:28 o/ catalyst 23:00:31 hi catalyst ! 23:00:56 gaba: our first item is roadmap, should we just bring in the part you have added directly here? 23:01:02 ahf: sort of 23:01:09 mikeperry: sweet 23:01:22 yes, I need you all to look at the roadmap and see if you are ok with everything there 23:01:43 i know teor has some things that they wanted to remove from the session we had a few weeks ago 23:01:47 maye other people too 23:02:14 our board is at https://dip.torproject.org/torproject/core/tor/-/boards 23:02:33 if you have any ticket on your plate that you are not going to do please release it 23:03:00 how does one "release" ? 23:03:13 desassign yourself from the ticket? 23:03:58 looks good to me at present... 23:04:04 probably going to move things around on monday... 23:04:09 ok 23:04:15 looks also good to me as well, except i need to update the title of #5304 which have gotten the title of a newer ticket 23:05:02 ok, cool, does it look ok to others too? 23:05:09 does #5304 really belong in needs_review? that's where it's lissted on dip.. 23:05:29 ah, nm, the title... 23:05:32 mikeperry: I do not have you in any ticket so far. You are going to create the tickets that are missing and I'm guessing you are going to collaborate with metrics team on the onionperf project. 23:05:32 now i understand 23:05:35 yeah, it is the title one 23:05:43 i don't have my 2fa device here so gonna rename it afterwards 23:05:57 nickm: it is quite possible that there are mistakes. please updae it if that is the case. 23:06:06 i will update it asap after the meeting 23:06:34 I am here, sorry, got caught up checking emails and tickets 23:06:35 gaba: yah, I think most of my work wrt this roadmap is actually metrics team work, except for this congestion control stuff I am just about to post 23:06:35 thanks 23:06:46 hi teor! 23:06:50 (and hi mikeperry !) 23:07:24 o/ teor 23:07:25 (hello world) 23:07:30 o/ arma2 23:08:25 hiya arma 23:08:49 are people OK with moving to reviewer assignments? i think neither david nor asn is here, but i checked earlier and it looked like things had been deligated out 23:08:56 anybody who wants to swap reviews with anybody else? 23:09:25 i take that as a no 23:09:34 i can take an extra review if need be: my current items are marked for 0.4.4 and stalled 23:09:35 let's do 0.4.3 status: https://trac.torproject.org/projects/tor/wiki/org/teams/NetworkTeam/CoreTorReleases/043Status 23:10:19 nickm: what are we fishing after with the 043 status today? do we want folks to look at the list and take things and/or remove things that we don't think are important and/or ? 23:10:52 So here's what I'd hope for today: everybody look at the list of "slated for removal" tickets if you haven't, and save any that are important. 23:11:08 Everybody look at the unassigned tickets at the top of the page if you haven't, and assign any that you really want to do to yourself. 23:11:25 ahf and I (and maybe gaba?) will work on assigning the ones that nobody jumps on 23:11:39 (That's my idea; sound okay?) 23:12:00 hm, a lot of the slated for removal hvae had very recent activity 23:12:06 but maybe that is some tagging that happened 23:12:45 ah, no, it's roadmap keyword that is added 23:13:04 I need to give up all of my 0.4.3 tickets, so I can focus on Sponsor 55 23:13:17 I'm writing proposals today and tomorrow, but I can do the ticket updates next week 23:13:32 * ahf moves #31009 and #5304 into 043-should 23:14:00 I put a list of tickets I think we should keep in the discussion section of the pad :-) 23:14:27 teor: I think we should all do some 0.4.3.x tickets, though it's fine if there are some you can't do. Maybe we can talk more about it? 23:15:03 teor: have you deassigned yourself already? otherwise i can do that so they go back in the pool ? 23:15:11 I've already done quite a few 0.4.3 tickets, I just can't prioritise the ones I have left right now 23:15:19 I have not un-assigned myself 23:15:35 oki, but the list you have made in the discussion items is complete? 23:16:22 No, that's the list of tickets I think we should keep 23:16:45 do they all have 043-must or 043-should? If not, let's add those 23:16:48 okay, so they should be 043-should and shot not be assigned to you? 23:16:55 s/shot/should/ 23:17:27 I'll go fix them now ;-) 23:17:36 thanks! 23:17:37 thx teor 23:18:17 are everybody else OK with the 043 status page or? 23:18:26 as per the things nickm just suggested we should do 23:18:27 okay for now. 23:18:36 oki! 23:18:45 arma2: this is also where you can kick in if you have some suggestions there :-) 23:18:49 let's go to discussions! 23:18:50 ahf: Let's plan to assign stuff on Monday before the meeting, based on what isn't taken? 23:19:00 wait, first let's do announcements! 23:19:14 nickm: yes! do we think it will take around 30 min'ish? 23:19:20 sure 23:19:21 then we can do it after our jitsi thingy? 23:19:28 maybe just as an extension to that 23:19:28 let's give it a try :) 23:19:32 sweet! 23:19:45 okay folks, remember we have our retrospective the 29th at 20 UTC! 23:20:02 there is already a pad for it for everybody to add things to if they need to remember anything :-) 23:20:13 ok, then discussions 23:20:23 the first item by gaba we handled during the roadmap part 23:20:26 the 2nd item is: 23:20:31 -- 0.4.3.x: any blockers? -- 23:20:32 * Any blockers for an 0.4.3.x release? 23:20:39 i have none 23:20:51 if anybody does, please let me know by the end of the meeting. Else I'll release tonight :) 23:20:54 and i think nickm have talked with dgoulet and asn about the OB-related changes, right? 23:21:11 onionbalance 23:21:24 what about that? 23:21:39 wasn't there a discussion on whether that was gonna be in 0.4.3 or in 0.4.4 ? 23:21:46 maybe i misread that conversation 23:21:53 right. We're targetting that branch for 0.4.4 so we can get it right 23:21:58 cool! 23:22:23 okay, the next item is added by teor and is about how we can all help with s55 23:22:46 teor: what is the discussion part here? i read it as some things that would make life easier to make progress quickly on this, right? 23:24:41 or does anybody have anything to add or discuss around the s55? 23:24:43 Since I'm the other person on S55, I'm guessing I should do a lot of the reviews. But we'll probably want an extra non-nickm reviewer on this. ahf: can you find somebody? 23:24:54 (We're doing a S55 planning meeting tomorow, right?) 23:25:19 i'd like to help out with reviews on the S55 tickets. i also had a conversation with teor about getting them access to a v6 relay they can run during this 23:26:46 (Also, can we annotate s28 and s30 with their general areas?) 23:26:55 annotate? 23:27:00 like, I am not so good at remembering numbers 23:27:11 s55 is ipv6, s27 is onion servies, s28 and s30 are ____ ? 23:27:11 annotate where? 23:27:20 s28 is pt and anti-censorship 23:28:10 and s30 is anticensorship too? 23:28:14 https://trac.torproject.org/projects/tor/wiki/org/sponsors/Sponsor30 and https://trac.torproject.org/projects/tor/wiki/org/sponsors/Sponsor28 23:28:33 Sorry, I was doing tickets 23:28:34 (thanks and sorry for the digression) 23:28:59 nickm: where do you want it annotated? on the pad or? 23:29:16 Yeah, I'm meant to be the only developer on Sponsor 55 (Relay IPv6 reachability and auto IPv6 address detection), and that was a bit of a surprise to me 23:29:17 teor: ah, cool! i have some questions above 23:29:42 ahf: i'm annotating on the pad, under the discussion/roadmap section 23:30:01 I'm happy to do most of the work, but I'm also happy for other people to pick up individual tickets 23:30:10 ah! 23:30:19 And you're right, ahf, let's have at least 2 reviewers 23:30:52 teor: was nickm's suggestion 8) but i think it is good as well. i will happily do reviews during the week too 23:31:04 teor: is there any of this that is also related to PTs and IPv6? 23:31:11 i know of one ticket that is in the intersection of those two 23:31:21 No, bridges and clients are out of scope. Public relays only. 23:31:32 We had to limit scope because the funding is limited. 23:31:45 Today and tomorrow, I'm working on the 3 proposals for sponsor 55: 23:32:00 Relay IPv6 Reachability & Relay IPv6 Extends 23:32:07 ah, oki 23:32:08 Relay Auto IPv6 Address Detection 23:32:19 Relay IPv6 Statistics 23:32:42 The network-team can help by reviewing those draft proposals, once I send them to tor-dev 23:32:51 cool 23:33:00 that sounds good 23:33:29 what time is the planning meeting again? 23:33:30 sounds good. I'll try to give you a one-working-day turnaround 23:33:43 23 hours after the start of this meeting 23:33:50 iiuc 23:34:04 ahf: you can look at the pad and add notes there 23:34:05 so at ... 22 UTC? 23:34:08 it may be hard for you to participate 23:34:20 Yeah, 2200 UTC Thursday 23:34:23 no, it's fine, i'm zombie all this week because of the construction 23:34:29 good, i have it as 22 UTC in my calendar already 23:34:32 cool! 23:35:02 And then once we have code, network team and relay operators can help with testing 23:35:03 teor: oki, next item we have is related to the 0.4.3 tickets, but that is what we did earlier in the meeting right and that you just deassigned on trac? 23:35:57 Yes, that's done, but one more thing: 23:36:04 If network-team people want to become master branch mergers, now would be a good time :-) 23:36:47 I think asn isn't doing as much merging lately, and so that just leaves nickm, dgoulet, and me, and 2 of us will be on sponsor 55 23:37:04 asn can't keyboard much right now 23:37:04 Merging is usually not a big task, but we should share it around 23:37:11 Yes I know ;-) 23:37:41 i'm considering it, but then i think it will be in early february, so i will get back to that 23:37:49 i do agree for master we should pass it around 23:38:40 ok, 20 min. left should we jump to the stem CI failure item? 23:39:17 yes 23:39:38 Are there a bunch of stem tests that time out, or just a couple? 23:39:46 And would it help if I picked up #30901? 23:39:56 Yes, #30901 would help 23:40:16 ok. 23:40:39 Diffing the stem trace log between failed and successful tests could also help, but you'd need to find the part of the log around the hang time 23:40:44 assignd to me as 043-should 23:40:50 At the moment, we just show the last 1000 lines 23:41:27 If there are just a small number of tests, I think that my CI wisdom says "we should disable the tests that aren't passing, and make our CI just run the reliable tests" 23:41:38 but if there are a lot of tests that time out, that's not an option 23:41:51 nickm: I have some incomplete code for #30901: ​https://github.com/torproject/tor/pull/1386 23:42:20 It seems that stem's testing framework / tor's control port is vulnerable to race conditions 23:43:10 Maybe if stem fixes the race condition in https://github.com/torproject/stem/issues/52 then other tests will start working 23:44:26 I hope that the tracing will help debug this 23:44:32 Let's check back once we've done #30901 ? 23:44:35 ok 23:45:06 cool 23:45:19 should we do the SBWS item? 23:46:08 makes sense. I also think we should pull an extra person into sbws so we have at least two staff on it 23:46:21 maybe gk is interested? It's sorta health. 23:46:34 I can also try to learn a little and do some; not sure how hard the tickets are 23:46:35 we are working on the network health roadmap tomorrow 23:47:00 i think having it as some form of split between network and network health could be smart 23:47:23 gaba and i will see juga next week, so hopefully we can learn what they are up to too 23:47:39 * gaba will add it to the pad for tomorrw - https://pad.riseup.net/p/tor-networkhealth-2020.1-keep 23:47:42 Yes, I think it's important we have 2 network team people on sbws while it's still in alpha and needs fixes 23:47:44 great; ty 23:47:45 And it's important we have someone from network health involved 23:48:57 teor: Please see my latest comment in https://github.com/torproject/stem/issues/52. That test is very simple, and should be reproducable without running the tests if you'd like. I don't think attempting to diff the test output is the best approach. 23:50:22 atagar: I'm sorry, but I just don't have time to work on this right now 23:50:29 can't this timeout value be bumped? we have seen things taking crazy long time on the appveyor and travis boxes when it has to do with scheduling 23:50:45 we had issues with process_t tests in the beginning where it could take 1m02s for it to be detected in the event loop 23:50:54 No, this issue is reproducible locally 23:51:01 ah! 23:51:06 And it only happens with --enable-fragile-hardening 23:51:07 teor: That's perfectly fine with me. What you were discussing above seemed incorrect, that's all. 23:51:12 On master, and not 0.4.2 23:52:32 okay, i do not see any other discussion items in the list 23:52:35 So it appears to be a timing race condition 23:52:42 does anybody have anything else we need to look at before we close off the meeting? 23:52:57 no, thanks ahf! 23:53:08 (gaba, ahf: I'd like to suggest my general idea of giving other teams some budget of points to ask for network team work on.) 23:53:21 (but that's not a must-do-at-this-meeting item) 23:53:22 nickm: that is a good idea 23:53:23 thanks, all! 23:53:30 a very good idea 23:53:32 yes, good idea. We can chat about it later 23:53:34 okay, i'm gonna call it off them 23:53:37 #endmeeting